Some of the most productive interactions in the workplace occur when a small group of people get together at a blackboard or a whiteboard and actively participate in presenting and discussing ideas. However it is very hard to support this style of interaction when participants are at different locations, a situation that occurs more and more frequently as organizations become more geographically distributed. To date, conventional video-conferencing systems are not well suited to this scenario. Effective collaboration relies on the ability for the parties to see each other and the shared collaboration surface, and to see where the others are looking and/or gesturing. Conventional video-conferencing systems can use multi-user screen-sharing applications to provide a shared workspace, but there is a disconnect from the images of the remote participants and the cursors moving over the shared application.
FIGS. 1-3 show schematic representations of systems configured to project images without interfering with images captured by a camera. FIG. 1 shows a communication medium with a half-silvered mirror 102, a camera 104 located above the mirror 102, and a projector 106. The mirror 102 and the projector 106 are positioned so that an image of a person or object located at a remote site is projected by the projector 106 onto the rear surface of the half-silvered mirror 102 and is visible to a participant 108. The camera 104 captures an image of the participant 101 via the participant's reflection in the mirror 102 and transmits the image to another person. The configuration of mirror 102, projector 106, and camera 104 enable the participant 108 to have a virtual face-to-face interaction with the other person. However, close interaction between the participant 108 and the other person can be disconcerting because the tilted screen makes for unnatural views of the remote user. FIG. 2 shows a communication medium with a switchable diffusing screen 202, a camera 204, and a projector 206. The screen 202 can be composed of a material that can be cycled rapidly between diffusive and transparent states. The state of the screen 202, projector 206, and camera 204 can be synchronized so that the projector 206 projects images when the screen is diffusive and the camera 204 captures images when the screen 202 in transparent. However, it is difficult to design a screen that can switch fast enough to avoid flicker, and the need to synchronize these fast switching components adds to the complexity of the system and limits the projected and captured light levels. FIG. 3 shows a top view of a communication medium with two cameras 302 and 304 on each side of a display 306. Images of a participant 308, for example, are captured by the cameras 302 and 304 and processed to create a single image of the participant 308 which appears to be captured by a single virtual camera 310 for viewing by another person at a different location. However, an image captured in this manner typically suffers from processing artifacts, especially when the captured views are at a very different angle from the intended virtual view, as would be the case with a participant located close to a large screen. This system also fails to capture hand gestures near, or drawing on, the screen surface.
It is desirable to have visual-collaborative systems that display images without interfering with and diminishing the quality of the images simultaneously captured by a camera.